DISCUSSION PAPER

The comparative economic statusof CDEP and non-CDEPcommunityresidents in the Northern Territoryin 1991

J.C. Altman and B. Hunter

No. 107/1996

ISSN 1036-1774ISBN 0 7315 1781 4

SERIES NOTE

The Centre for Aboriginal Economic Policy Research (CAEPR) wasestablished in March 1990 under an agreement between The AustralianNational University (ANU) and the Aboriginal and Torres Strait IslanderCommission (ATSIC). CAEPR operates as an independent research unitwithin the University's Faculty of Arts and is funded by ATSIC, theCommonwealth Department of Social Security and the ANU. CAEPR'sprincipal objectives are to undertake research to:

• investigate the stimulation of Aboriginal and Torres Strait Islandereconomic development and issues relating to Aboriginal and TorresStrait Islander employment and unemployment;

• identify and analyse the factors affecting Aboriginal and Torres StraitIslander participation in the labour force; and

• assist in the development of government strategies aimed at raisingthe level of Aboriginal and Torres Strait Islander participation in thelabour market.

The Director of the Centre is responsible to the Vice-Chancellor of theANU and receives assistance in formulating the Centre's research prioritiesfrom an Advisory Committee consisting of five senior academicsnominated by the Vice-Chancellor and four representatives nominated byATSIC, the Department of Employment, Education, Training and YouthAffairs and the Department of Social Security.

CAEPR Discussion Papers are intended as a forum for the rapiddissemination of refereed papers on research that falls within the CAEPRambit. These papers are produced for discussion and comment within theresearch community and Aboriginal affairs policy arena. Many aresubsequently published in academic journals. Copies of discussion paperscan be purchased from:

Centre for Aboriginal Economic Policy Research,Faculty of Arts, The Australian National University, Canberra ACT 0200.Ph (06) 279 8211 Fax (06) 249 2789.

Abstracts of all CAEPR Discussion Papers can be found at the followingWorld Wide Web address:

http://coombs.anu.edu.auAVWWVLPages/AborigPages/CAEPR/caepr-home.html

As with all CAEPR publications, the views expressed inthis Discussion Paper are those of the author(s) and do

not reflect an official CAEPR position.

Professor Jon AltmanDirector, CAEPRThe Australian National University

ABSTRACT

The Community Development Employment Projects (CDEP) scheme is aprogram where participants forego social security entitlements and workfor rough equivalents of these entitlements. The scheme has expandedrapidly over the past decade and now operates in 250 Indigenouscommunities. While the scheme has multiple objectives, it is arguablyprimarily a labour market program. While there is a growing body ofresearch about the scheme, key research and policy questions about thelabour market effects of the scheme on participating communities, incontrast to other communities, have never been asked.

This paper represents the first attempt to compare labour marketoutcomes at a sample of CDEP and non-CDEP communities in theNorthern Territory. This exercise is undertaken with community profiledata (on Indigenous people only) from the 1991 Census: ten CDEP andnine nearby non-CDEP communities are compared. Five variables -labour force status, industry structure, occupational status, sector ofemployment and income status - are analysed.

The paper's results are potentially of considerable policy significance.While official employment levels at CDEP communities are predictablyhigher than at non-CDEP communities, these levels exceed the directeffect of scheme participation. However, income status at CDEPcommunities is a little lower than at non-CDEP communities, a puzzlingoutcome given the options to earn additional income when participating inthe scheme. Another puzzling outcome, counter to the views of someresearchers, is that there is little statistical evidence of the scheme beingused as a substitution funding regime. It is recommended that results aretreated with caution given the regional focus of this exploratory analysis.

Acknowledgments

An early version of this paper 'Discernible impacts of the CDEP scheme:census-based case studies from the Northern Territory' was presented as aCAEPR seminar in September 1993. The paper was revisited when DrBoyd Hunter and Ms Liu Jin recalculated data for selected communities.We would like to thank all participants at the 1993 seminar and, inparticular, Max Neutze, whose comments were extremely valuable.Thanks also to John Taylor and Will Sanders for comments; to Liu Jin fordata manipulation and cartography; to Hilary Bek for research assistance;and to Linda Roach and Joy Humphreys for production assistance.

Professor Jon Altman is Director and Dr Boyd Hunter is Post-DoctoralFellow at the Centre for Aboriginal Economic Policy Research, Facultyof Arts, The Australian National University, Canberra.

The Community DevelopmentEmployment Projects (CDEP) scheme is theAboriginal and Torres Strait Islander Commission's (ATSIC's) largestprogram, a program that is regarded primarily as a labour market program.1

In 1996, twenty years after its introduction as a pilot, the CDEP schemeremains somewhat ambiguous and contested. In recent years, and inparticular since the launch of the Aboriginal Employment DevelopmentPolicy (AEDP) in 1987, considerable research has been undertaken on thescheme.2 Much of this research has highlighted the need for statisticalinformation about the effectiveness of the scheme. A particular problemhas been the absence of a direct means to identify CDEP schemeparticipants from census data (see Altman and Daly 1992). It is likely thatthis shortcoming will be rectified in coding of 1996 Census data. Therecently completed National Aboriginal and Torres Strait Islander Survey(NATSIS) also included a specific question about CDEP employment (seeAustralian Bureau of Statistics (ABS) 1995; ABS/CAEPR 1996).

A fundamental research and policy question, however, has not been askedabout the scheme, namely how have groups and communities who havejoined the scheme fared in comparison with other groups who have not.This paper, somewhat belatedly, utilises 1991 Census data to address theimportant policy research question: do the socioeconomic characteristics ofresidents of communities which participate in the CDEP scheme differ inany significant or discernible way from those of residents of communitieswho do not participate in the scheme? When unit record data from NATSISbecome available later in 1996, such an issue can be addressed forindividuals and, when 1996 Census data become available late in 1997,such a question could be effectively answered as all CDEP schemeparticipants will be identified. Until then, only community profiles data,combined with administrative information, can be used to contrast keyeconomic indicators at communities participating in the scheme with othercommunities which, in 1991, remained outside the scheme's coverage.

Methodology

The primary focus of this paper is to examine, using 1991 Census data, thesimilarities and differences between a select sample of ten communitiesparticipating in the CDEP scheme with nine that are not. The paper focuseson the Northern Territory (NT) for three reasons. First and foremost, anumber of discrete CDEP scheme and non-CDEP scheme communities(hereafter abbreviated to CDEP and non-CDEP communities) can beidentified in published community profile data (ABS 1993). Second, thefact that these communities are relatively geographically discrete meansthat to a great extent ABS census and ATSIC administrative data sets couldbe correlated. Third, the CDEP scheme has been in existence in the NT aslong as anywhere in Australia.

Communities selectedIn 1991, according to ATSIC administrative data sets, there were 28communities or outstation resource centres in the NT that participated inthe CDEP scheme; there were 4,146 individuals participating in thescheme. For this analysis, ten of these communities were selected becausethey formed discrete identifiable participating communities. A decisionwas made to avoid communities with associated outstations participating inthe scheme as this would have made correlation of census geography(community profiles rather than regional data) and administrative dataincompatible (Altman and Daly 1992).

The CDEP communities chosen for analysis were Beswick (Wugularr),Gapuwiyak (Lake Evella), Lajamanu, Milikapiti, Ngukurr, Pularumpi, PineCreek (Kybrook Farm), Santa Teresa, Umbakumba and Willowra. Thegeographic distribution of these communities is indicated in Figure 1. Awide geographic range, within the constraints outlined above, wasintentionally sought. In 1991, there were 1,381 CDEP scheme participantsat these ten communities, representing 33 per cent of all NT participants.

Figure 1. Location of CDEP and non-CDEP communities in the NT.

APularmBgi A Milikapiti

ANgmu

J A Gapuwiyak

Ramingin_ gtf? ^Angurugu

APine Creek ~ ""y^AUmbakumbaABeswick J *Ja

A NgiAJcurKatherine

A Lajamanu

Borroloola

AWillowra

YuendumuATI Tree

A Amoonguna

ASanta TeresaACDEP communitiesANon-CDEP communities

The nine communities that were chosen as the 'control' group because theywere not participating in the scheme were selected in an attempt to achievea similar geographic spread as for participating communities. Thesecommunities, also indicated in Figure 1, were Amoonguna,Angurugu,Borroloola, Bulman, Katherine township, Nguiu, Ramingining, Ti Tree andYuendumu. For both CDEP and non-CDEP communities an attempt wasmade to select communities that are isolated from one-off influences or, ifa major development was influential, then we ensured that both a CDEPand non-CDEP community from the region were included in our sample.3

Table 1. Some population and labour force characteristics at selectedCDEP and non-CDEP communities, 1991.

Community15

BeswickWillowraUmbakumbaSanta TeresaPularumpiPine CreekNgukurrMilikapitiLajamanuGapuwiyakTotal

BulmanTi TreeAnguruguAmoongunaNguiuKatherineBorroloolaYuendumuRaminginingTotal

Populationyears plus

18818918923615694

429266365180

2,292

59217378101615869195347242

3,023

Labour Participation CDEP scheme Year joinedforce rate participants scheme

1281221001158548

28020322743

1,351

12616333

12638912213388

1,027

68.164.652.948.754.551.165.376.362.223.958.9

20.328.116.732.720.544.862.638.336.434.0

128130104876462

30815025098

1,381

1988-891988

1990-911986-871989-901989-901988-891988-891989-90

1980

Sources: ABS (1993); ATSIC administrative data.

Table 1 indicates communities chosen, populations of working age (agedover 15 years), labour force participation rates, CDEP scheme participantnumbers (from ATSIC data sets, at July 1991) and the year eachcommunity joined the scheme. Despite the attempt to select similarcommunities, CDEP communities are, on average, smaller (357 persons)

than non-CDEP communities (555 persons). Two other features of Table 1are noteworthy. First, all CDEP scheme communities, except for'Gapuwiyak, only began participating in the scheme since 1988. Thisreflects, in large part, the rapid expansion of the scheme since the launch ofthe AEDP in 1987 (Taylor 1993a). Indeed, in 1986, there were only fourcommunities with 720 participants in the scheme in the NT. Second, it isimportant to recognise that not all scheme participants are in the labourforce: despite the fact that the CDEP scheme is often referred to as a 'work-for-the-dole' scheme, not all participants work. Hence, in Table 1, it isevident that at some communities the number of CDEP scheme participantsexceeds the size of the labour force. This is an issue that is discussed ingreater detail below.

CaveatsIt is important to note that the analysis in this paper is circumscribed byseveral caveats most of which arise from the fact that CDEP schemeparticipants were not identified in the 1991 Census. The communities inthis analysis are selected, rather than randomly chosen, primarily becausethey cover a number of communities known to either participate or notparticipate in the scheme. Because the sample represents a significantproportion of the adult Indigenous population (23 per cent) and labourforce (24 per cent) in the NT, this is not a major shortcoming but,nevertheless, it should be borne in mind. In particular, communities werechosen in an effort to ensure that no major structural differences existedbetween the two types (CDEP and non-CDEP) of communities.

Second, there are data shortcomings. While the analysis here focuses oncommunity profile data from the 1991 Census, in places use is made ofATSIC administrative data sets (CDEP participant numbers). It should benoted that these two data sets are collected by very different methods. Thiscreates problems in attempts to reconcile the number of CDEP participants(1,381 in Table 1) with the size of the labour force (1,351) at CDEPcommunities. It also creates uncertainty about the allocation of CDEPparticipants between employed, unemployed and not in the labour force.This reconciliation problem is addressed in Appendix 1. It is important torecognise that not all part-time employment in CDEP scheme communitiesoccurs under the auspices of the scheme. It is impossible to assess fromcensus data if employment is generated by the CDEP scheme; nevertheless,it must be recognised that some part-time employment is available forIndigenous people at all CDEP communities that is quite independent ofthe scheme. Finally, in Tables 4 to 7 the 'not stated' category is very high,particularly at non-CDEP communities. It is difficult to find reasons forthis, but it highlights that data must be treated with caution.

General findings

The data from community profiles have been tabulated into six tables thatdifferentiate a number of variables at CDEP and non-CDEP communities.Particular attention is focused on employment status, industry structure,occupational structure, sector of employment and income status.

Employment and unemploymentTable 2 presents data on population, employment and unemployment, andemployment/population and unemployment/population ratios, in CDEPand non-CDEP communities. In aggregate terms, it is apparent that theemployment/population ratio at CDEP communities at 56.3 per cent ismuch higher than at non-CDEP communities (at 19.9 per cent).Conversely, the unemployment/population ratio (2.7 per cent) is muchlower at CDEP communities than at non-CDEP communities (14.0 percent). At CDEP communities there is one anomalous result: Gapuwiyakhas an unusually low employment/population ratio of 22.2 per cent.

Table 2. Employment and unemployment at selected CDEP and non-CDEP communities, 1991.

Population Employed Employment/ Unemployed Unemployment/over 1 5 years population ratio population ratio

BeswickWillowraUmbakumbaSanta TeresaPularumpiPine CreekNgukurrMilikapitiLajamanuGapuwiyakTotal

BulmanTi TreeAnguruguAmoongunaNguiuKatherineBorroloolaYuendumuRaminginingTotal

18818918923615694

429266365180

2,292

59217378101615869195347242

3,023

12511397

1157645

28019420540

1,290

9215727

120230

415642

603

66.559.851.348.748.747.965.372.956.222.256.3

15.39.7

15.126.719.526.521.016.117.419.9

39309309

223

61

340666

159817746

424

1.64.81.60.05.83.20.03.46.01.72.7

5.118.4

1.65.91.0

18.341.522.219.014.0

Source: ABS (1993).

If 'non-CDEP' employment is factored out (making the assumption that thenon-CDEP communities' employment/population ratio of 19.9 per centapplies to CDEP communities), then it can be calculated that 834 personsat CDEP communities are 'employed1 because of the scheme. Thisrepresents 60 per cent of scheme participants at these communities (seeTable 1) according to ATSIC administrative data, a figure that is the sameas an estimate made by Taylor (1993a, 1993b) based on case study dataprovided by Deloitte Touche Tohmatsu (1993).4 Other options fordistributing CDEP scheme participants are examined in Appendix 1. Anadditional observation is that the existence of the CDEP scheme in acommunity should preclude the possibility of unemployment, at least as faras eligibility to Jobsearch Allowance or Newstart is concerned. By andlarge this is the case, although at some CDEP communities up to 5 or 6 percent of the population was classified, or identified themselves, asunemployed for census purposes.

Labour force statusIn Table 3, summary information is presented on the overall labour forcestatus of residents of CDEP and non-CDEP communities. This informationfocuses more closely on the nature of employment, particularly withrespect to hours worked. It is no surprise that a significant proportion (31.5per cent) of the adult population at CDEP communities is employed part-time: the key feature of the scheme is the provision of part-time work forthe equivalent of social security entitlements.

Table 3. Labour force status of residents of CDEP and non-CDEPcommunities, 1991.

Labour forcestatus

EmployedPart-timeFull-timeNot stated

Total employedUnemployedNot in labour forceTotal population over 15

CDEP communitiesNumbers

72352740

1,29061

9412,292

Per cent

31.523.0

1.756.3

2.741.1

100.0

Non-CDEP communitiesNumbers

229296

78603424

19963,023

Per cent

7.69.82.6

19.914.066.0

100.0

Source: ABS (1993).

What is surprising though is that full-time employment is more prevalent,in absolute terms, in CDEP communities than at non-CDEP communities

and that a significant proportion of the adult population (7.6 per cent) atnon-CDEP communities is employed part-time. These data can be variablyinterpreted. Firstly, it is possible that some CDEP part-time work is beingconverted to full-time work via the utilisation of a mix of CDEP and othersources (Smith 1994). Using the estimate of 834 CDEP jobs atparticipating communities and assuming a similar proportion of non-CDEPpart-time employment (174 jobs), then a large number of CDEP jobs (285out of 834, or 34 per cent) could be classified in the census as full-time.

Secondly, it has been suggested in the literature that the CDEP schemeoperates as a substitution funding regime, with State and local governmentsallowing the scheme to fund employment of Indigenous people in areasthat would normally be financed by government (Altman and Sanders1991; Deloitte Touche Tohmatsu 1993). This is an issue that will beexamined further below, but there is no clear empirical evidence ofsubstitution in Table 3 given that full-time employment levels are higher inCDEP than non-CDEP communities.

It is especially significant that labour force participation is significantlyhigher in CDEP communities (59 per cent) than in non-CDEP communities(34 per cent), assuming of course that CDEP work is accepted as realemployment. Obviously a high proportion of the gap is made up of thoseCDEP scheme participants who are joining the workforce.

Industry of employmentInformation is provided in Table 4 on industry of employment. There areonly a few interesting differences between CDEP and non-CDEPcommunities. Nearly all employment in CDEP communities (87 per cent)is in two industries - community services and public administration -compared with 47 per cent in these two industries in non-CDEPcommunities. Overall, in non-CDEP communities, industry of employmentis more diverse, with wholesale and retail trade standing out as an area ofdifference.

The utility of the classification of CDEP workers in the census can bequestioned, if only because it varies so significantly from informationcollected by ATSIC at CDEP communities. For example, Taylor (1995:10) presents information from the ATSIC census of CDEP schemeactivities covering approximately one-third of scheme participants in theNT in December 1994. He found a far wider spread of activities, withsignificant CDEP employment in areas such as agriculture, construction,arts and crafts, and recreation and personal service industries. This findingis consistent with that of the Deloitte Touche Tohmatsu (1993) study thatalso found a diversity of activities in which CDEP scheme workers engage.This discrepancy arises because the five-yearly census has classified CDEPscheme participants into the community services and public administration

industries of employment on the basis of their employer (local governmentor community council) rather than by examining the nature of the activitythey undertake (Taylor 1995: 9-10). '

Table 4. Industry structure in CDEP and non-CDEP communities,1991.

IndustryCDEP communities Non-CDEP communities

Numbers Percent Numbers Percent

Agriculture, forestry, fisheriesand hunting 3 0.2 18 3.1Mining 0 0.0 3 0.5Manufacturing 12 0.9 21 3.6Electricity, gas and water 6 0.5 0 0.0Construction 27 2.1 19 3.3Wholesale and retail trade 33 2.6 100 17.3Transport and storage 6 0.5 6 1.0Communication 0 0.0 3 0.5Finance, property and business services 0 0.0 10 1.7Public administration and defence 381 29.6 90 15.5Community services 737 57.3 183 31.6Recreation and personal services 9 0.7 22 3.8Not classifiable 0 0.0 0 0.0Not stated 72 5.6 104 18.0

Total 1,286 100.0 579 100.0

Source: ABS (1993).

Occupational statusIn Table 5, information is provided on occupational status. It should benoted that the total number of persons reporting an occupation in CDEPcommunities exceeds the total number employed. This could be due toCDEP scheme participants identifying an occupation for themselves butnot considering themselves employed.

The outstanding feature of Table 5 is that a very high proportion (44 percent) of those employed at CDEP communities were classified in thelabourer category, supporting a view that the CDEP scheme only generatesopportunities for low-skilled employment without appropriate training andexit options. Interestingly, while the proportion in many occupationalcategories is lower in CDEP communities, this merely reflects higheroverall employment levels. In absolute terms the numbers in higher skilledoccupations like professional, paraprofessional and the trades were higherin CDEP communities, suggesting that the existence of the scheme may in

fact generate additional opportunities. This calls into question anecdotalviews that skilled occupational opportunities and training opportunities aredeficient at CDEP communities.

Table 5. Occupational structure of CDEP and non-CDEPcommunities, 1991.

CDEP communities Non-CDEP communitiesOccupation Numbers Percent Numbers Percent

Managers and administratorsProfessionalsParaprofessionalsTradespersonsClerksSales and personal service workersPlant and machine operatorsLabourers and related workersInadequately describedNot stated

307190

15789

12861

5824079

2.35.46.8

11.86.79.64.6

43.93.06.0

30673971967226

1101280

5.011.16.5

11.815.911.94.3

18.22.0

13.3

Total 1,327 100.0 603 100.0

Source: ABS (1993).

Sector of employmentTable 6 provides data on industry sector of employment. The outstandingfeature of this table is evidence that a very significant share of employmentat CDEP communities is in the local government sector (55 per cent).Indeed, in absolute terms local government employment in CDEPcommunities exceeds all employment in non-CDEP communities.However, there is a possibility that some of this concentration may be dueto the fact that incorporated community councils (local governments) areresponsible for administrating the CDEP scheme.

Interestingly though, NT Government employment is more prevalent inboth absolute and population-relative terms in the CDEP communities,while Commonwealth government employment is lower. This finding ispotentially significant because it challenges the view that State/Territorygovernments withdraw support from communities participating in theCDEP scheme. Certainly in the NT in 1991 this does not appear to havebeen the case. Private sector employment is similarly more significant atCDEP communities. It is important to note though that there is a tendencyfor private sector (or non-government) employment to be overstated at all

10

Indigenous communities owing to significant employment, includingCDEP scheme employment, in the community sector.

Table 6. Sector of employment in CDEP and non-CDEP communities,1991.

CDEP communities Non-CDEP communitiesSector Numbers Per cent Numbers Per cent

CommonwealthState/TerritoryLocalPrivateNot stated

1711670036680

1.39.1

54.728.66.3

3511268

28792

5.918.911.448.315.5

Total 1,279 100.0 594 100.0

Source: ABS (1993).

Income statusIn Table 7, information is provided about the incomes of individualsresiding at CDEP and non-CDEP communities. This table provides somesignificant findings. Median individual income is $39 per annum higher atnon-CDEP communities, while mean income is $457 higher. While themore substantial mean income difference is not significantly different, thesource of this difference is of interest. Higher income at non-CDEPcommunities is largely driven by the higher proportion of people earningover $16,000 per annum at these communities, and particularly over$20,000 per annum. Given the greater incidence of full-time work in CDEPcommunities (Table 3) this suggests that full-time work at non-CDEPcommunities is better paid than at CDEP communities. The difference inemployment levels at CDEP communities noted above is not reflected inhigher incomes.

This finding is consistent with previous reports of intercensal (1986 to1991) income stagnation at Aboriginal communities (Taylor 1993a, 1993b;ATSIC 1994; Altman and Daly 1995). At one level, it is predictable thatparticipation in a work-for-the-dole scheme will only provide wagesequivalent to social security entitlements. At another level though, it issurprising that CDEP scheme participation does not generate additionalincome, particularly in those situations where it is used to facilitateparticipation in income-generating activities like production of artefacts forsale because income testing in CDEP communities is more generous than

II

at non-CDEP communities. Taylor (1995: 10) provides some informationon the range of activities that over 2,000 participants in the scheme in theNT undertook in December 1994. While this information was from anATSIC survey rather than from the ABS, it still included a number ofactivities that were potentially income-generating. It is also noteworthy thatall CDEP communities receive additional resources beyond notionalwelfare entitlements to meet administrative on-costs (that often create jobs)and to provide some capital support for income-generatingactivities.

Table 7. Income status at CDEP and non-CDEP communities, 1991.

Income ($)CDEP communities Non-CDEP communities

Numbers Per cent Numbers Per cent

0-3,0003,001-5,0005,001-8,0008,001-12,00012,001-16,00016,001-20,00020,001-25,00025,001-30,00030,001-40,00040,000+Not statedTotal

MeanStandard error of meanMedian

146215764549218753733126

2292,284

8,900237

7,617

6.49.4

33.524.09.53.31.61.40.50.3

10.0100.0

30322580354226715811139246

5453,023

9,357238

7,656

10.07.4

26.617.98.85.23.71.30.80.2

18.0100.0

Source: ABS (1993).

It remains unclear why these additional resources are not generating higherindividual incomes at CDEP communities. One possibility is that there issystematic under-reporting of additional informal income earned at CDEPcommunities or that additional productive work is generating non-cashreturns in the form of unmarketed goods and services. Another possibilityis that in 1991 the scheme was not well enough established in the selectedcommunities to generate additional income.

12

Some policy implications

A major finding from this comparative analysis is that participation in theCDEP scheme moves many people from being located outside the labourforce to labour force participation and from unemployment to employment.This is a mixed result. If participants exit the CDEP scheme intomainstream employment in the longer term, then the scheme could be seento have operated as an appropriate vehicle to shift from being outside thelabour force to genuine employment. If, on the other hand, individuals aremoving from welfare-dependent 'not in the labour force' and unemployedstatus to CDEP-dependent part-time employment, then the value of thescheme, in labour economic terms, will be questioned (Verucci 1995).

It is frequently overlooked today that the influential Review of AboriginalEmployment and Training Programs (Miller 1985) identified the CDEPscheme as a potential income-generation mechanism particularly forremote communities (see Altman and Taylor 1989). Given thatparticipation in the scheme provides opportunities to earn more income,before income withdrawal occurs, than from social security, it is surprisingthat official statistics indicate that individual incomes at CDEPcommunities are little different (being a little lower) from those at non-CDEP communities. One would expect higher incomes at CDEPcommunities because employment levels suggest that people are workingmore, in addition to the part-time work available under the scheme.5

The policy challenge, in the current political environment with anincreasing emphasis on outcomes, is to demonstrate the relative benefits ofthe scheme given its additional financial cost to government, estimated at37 per cent more than welfare entitlements (ATSIC 1995: 70). Thechallenge for both ATSIC and participating communities will be to developrigorous methods to demonstrate outcomes, more broadly if necessary, toinsulate the scheme from unwarranted criticism. As noted at the start ofthis paper, the scheme remains both ambiguous and contested; a clearerdefinition of its goals and the collection and analysis of quantitative datathat demonstrate its contributions might ameliorate the lack of policyclarity about the scheme in future years. If the potential to exit the schemeis deemed a priority, then longitudinal research that replicates thecomparative analysis undertaken here will be essential.

Conclusion

This paper represents the first exploratory attempt to compare and contrastthe labour market outcomes for a number of communities participating inthe CDEP scheme with a control group. As such, it provides a somewhatdifferent angle on evaluating the effectiveness of the CDEP scheme. The

13

conclusions that can be drawn from this analysis, given its regional focusand the nature of sample selection, must be cautiously interpreted.Furthermore, while the analysis is intentionally empirical and quantitative,it may have overlooked both positive and negative qualitative aspects ofthe scheme that have been discussed in more detailed case studies (Deloittelouche Tohmatsu 1993; Smith 1994, 1995).

Even with these qualifications, the results of the comparative exerciseundertaken here are mixed. There is clear evidence that employment levelsare higher at communities participating in the CDEP scheme. Thisimprovement exceeds the direct employment created by the scheme andgoes beyond the cosmetic appearance of employment creation that occurswhen individuals move from welfare to workfare. On the other hand, thereis also evidence that income levels at CDEP communities are similar tothose at non-CDEP communities and may be slightly lower. This resultsuggests that in terms of policy goals to improve the economic status ofIndigenous Australians, the CDEP may be less effective than might beexpected. Consideration needs to be given to assessing whether this findingis widespread and, if this is the case, what measures can be introduced toensure that the CDEP scheme results in improved income status.

Notes

1. By 1995-96, the scheme had expanded to include 252 communities, with 27,041participants and costing $278.3 million (ATSIC 1995: 70).

2. This has included Sanders (1988), Altman and Sanders (1991) and Sanders (1993)on public policy aspects of the scheme; Altman and Daly (1992) and Verucci(1995) on labour economics aspects of the scheme; and Taylor (1993a, 1993b),ATSIC (1994) and ABS/CAEPR (1996) on economic policy aspects. There havebeen few detailed case studies about the scheme's operations, the majorexceptions being a study by Deloitte Touche Tohmatsu (1993) and Smith (1994,1995). A very comprehensive annotated bibliography is provided in an appendixin the Deloitte Touche Tohmatsu (1993) study.

3. Of particular relevance in 1991 was the existence of a major manganese mine onGroote Eylandt (where CDEP community Umbakumba and non-CDEP Anguruguare located) and tourism on Melville and Bathurst Islands (where non-CDEPNguiu and CDEP Milikapiti and Pularumpi are located).

4. This is an important verification of an earlier ballpark estimate that has been usedby ATSIC (1994) in its review of the AEDP. It is also similar to the total numberof people employed in the CDEP scheme as reported in NATSIS (16,800) as aproportion of CDEP scheme participants in 1994 (25,000).

5. However, while the CDEP scheme may not ameliorate welfare dependence, it hasa positive impact on employment activity at communities. That is, it may notincrease income but it could increase the level of services and therefore thequality of life for residents of CDEP communities. Therefore, income status andwelfare dependency may not be the sole measures of success or otherwise of thescheme.

14

Appendix 1. Options for reconciling ABS census and ATSIC CDEP schemeparticipation data.

An inconsistency between the size of the labour force at CDEP communities (1,351persons) and administrative data on participants (1,381) is evident in Table 1. Clearly,some CDEP participants (at least 30) are not in the labour force. However, it is alsoclear that there is employment at CDEP communities that is outside the ambit of thescheme. Because there is no identification of CDEP participants in the census, there isno definitive evidence to unequivocally support this view. However, there is indirectevidence like significant NT Government employment in Table 6 and significant full-time employment in Table 3.

In this discussion paper, it is assumed that non-CDEP employment at CDEPcommunities replicated levels at non-CDEP communities. Arguably, this assumes nosubstitution between NT Government employment at CDEP communities and CDEPscheme employment. This view is largely based on Table 6 because proportionallythere are more 'State' government jobs at CDEP communities than non-CDEPcommunities. These data could reflect a classification error in the 1991 Census, butfieldwork experience at a number of CDEP communities suggests that substitution isnot occurring with respect to NT Government responsibilities such as health, educationand community government services, at least not in contrast to non-CDEPcommunities. This does not preclude the possibility that both CDEP and non-CDEPcommunities are under-resourced by government, merely that there is no evidentstatistical difference between them.

If the assumption about the underlying structure of CDEP and non-CDEP communitiesbeing similar is accepted, it is calculated that the employment impact of the CDEPscheme is 834 jobs (according to the census) out of 1,381 participants (according toATSIC administrative data based on participant schedules). This raises questions aboutthe labour market classification of the other 547 CDEP scheme participants. At CDEPcommunities, 61 persons were classified as unemployed. If we accept that all theseunemployed are scheme participants, then there are 486 remaining participants toreconcile.

Three possibilities need to be considered. First, the census may have underenumeratedthe population, labour force, employment and unemployment at CDEP communities. Itis estimated that the 1991 Census underenumerated Indigenous Australians by 5 percent for females and 11 per cent for males (ABS/CAEPR 1996). Second, ATSICparticipant schedules may have overcounted participants. Third, CDEP schemeparticipants, especially non-working spouses of working participants receiving incomesupport from the scheme, may have been classified as 'not in the labour force'. All threeare possibilities. The first would have accounted for about one-third of the discrepancy.The overcounting of participants, if matched by payments, would have in all probabilityincreased individual incomes at CDEP communities. There is no evidence of an incomedifferential between CDEP and non-CDEP communities in Table 7. The thirdpossibility could have accounted for the balance.

References

Aboriginal and Torres Strait Islander Commission (ATSIC) 1994. Review of theAboriginal Employment Development Policy, ATSIC, Canberra.

Aboriginal and Torres Strait Islander Commission (ATSIC) 1995. Annual Report 1994-95, ATSIC, Canberra.

15

Altman, J.C. and Daly, A.E. 1992. The CDEP scheme: a census-based analysis of thelabour market status of participants in 1986', CAEPR Discussion Paper No. 36,Centre for Aboriginal Economic Policy Research, The Australian NationalUniversity, Canberra.

Altman, J.C. and Daly, A.E. 1995. 'Indigenous Australians in the labour market:historical trends and future prospects', Economic Papers, 14 (4): 64-73.

Altman, J.C. and Sanders, W. 1991. The CDEP scheme: administrative and policyissues', Australian Journal of Public Administration, 50 (4): 515-25.

Altman, J.C. and Taylor, L. 1989. The Economic Viability of Aboriginal Outstationsand Homelands, Australian Government Publishing Service, Canberra.

Australian Bureau of Statistics (ABS) 1993. 1991 Census of Population and Housing:Aboriginal Community Profile, cat. no. 2722.2, ABS, Canberra.

Australian Bureau of Statistics (ABS) 1995. National Aboriginal and Torres StraitIslander Survey 1994: Detailed Findings, cat. no. 4190.0, ABS, Canberra.

Australian Bureau of Statistics (ABS) and Centre for Aboriginal Economic PolicyResearch (CAEPR) 1996. National Aboriginal and Torres Strait Islander Survey1994: Employment Outcomes for Australia's Indigenous Peoples, cat. no. 4199.0,ABS, Canberra.

Deloitte Touche Tohmatsu 1993. No Reverse Gear: A National Review of theCommunity Development Employment Projects Scheme, Report to the Aboriginaland Torres Strait Islander Commission, Canberra.

Miller, M. (Chairman) 1985. Report of the Committee of Review of AboriginalEmployment and Training Programs, Australian Government Publishing Service,Canberra.

Sanders, W. 1988. The CDEP scheme: bureaucratic politics, remote communitypolitics, and the development of an Aboriginal 'workfare' program in times of risingunemployment1, Politics, 23 (1): 32-47.

Sanders, W. 1993. The rise and rise of the CDEP scheme: an Aboriginal 'workfare'program in times of persistent unemployment', CAEPR Discussion Paper No. 54,Centre for Aboriginal Economic Policy Research, The Australian NationalUniversity, Canberra.

Smith, D.E. 1994. '"Working for CDEP": a case study of the Community DevelopmentEmployment Projects scheme in Port Lincoln, South Australia', CAEPR DiscussionPaper No. 75, Centre for Aboriginal Economic Policy Research, The AustralianNational University, Canberra.

Smith, D.E. 1995. 'Redfern works: the policy and community challenges of an urbanCDEP scheme', CAEPR Discussion Paper No. 99, Centre for Aboriginal EconomicPolicy Research, The Australian National University, Canberra.

Taylor, J. 1993a. The Relative Economic Status of Indigenous Australians, 1986-91,Research Monograph No. 5, Centre for Aboriginal Economic Policy Research, TheAustralian National University,Canberra.

Taylor, J. 1993b. Regional Change in the Economic Status of Indigenous Australians,1986-91, Research Monograph No. 6, Centre for Aboriginal Economic PolicyResearch, The Australian National University, Canberra.

16

Taylor, J. 1995. 'Indigenous employment and job segregation in the Northern Territorylabour market', CAEPR Discussion Paper No. 83, Centre for Aboriginal EconomicPolicy Research, The Australian National University, Canberra.

Verucci, N. 1995. The Community Development Employment Projects Scheme; realemployment or disguised welfare', Economic Papers, 14 (4): 74-82.

CENTRE FOR ABORIGINAL ECONOMIC POLICY RESEARCH(CAEPR)

MONOGRAPHS

1. Aborigines in the Economy: A Select Annotated Bibliography of Policy-RelevantResearch 1985-90, L.M. Allen, J.C. Altman and E. Owen (with assistance fromW.S.Arthur), 1991.

2. Aboriginal Employment Equity by the Year 2000, J.C. Altman (ed.), publishedfor the Academy of the Social Sciences in Australia, 1991.

3. A National Survey of Indigenous Australians: Options and Implications,J.C. Altman (ed.), 1992.

4. Indigenous Australians in the Economy: Abstracts of Research, 1991-92,L.M. Roach and K.A. Probst, 1993.

5. The Relative Economic Status of Indigenous Australians, 1986-91, J. Taylor,1993.

6. Regional Change in the Economic Status of Indigenous Australians, 1986-91,J. Taylor, 1993.

7. Mabo and Native Title: Origins and Institutional Implications, W. Sanders(ed.), 1994.

8. The Housing Need of Indigenous Australians, 1991, R. Jones, 1994.

9. Indigenous Australians in the Economy: Abstracts of Research, 1993-94,L.M. Roach and HJ. Bek, 1995.

10. The Native Title Era: Emerging Issues for Research, Policy and Practice,J. Finlayson and D.E. Smith (eds), 1995.

For information on earlier CAEPR Discussion Papers please contact PublicationSales, Centre for Aboriginal Economic Policy Research, Faculty of Arts, AustralianNational University, Canberra ACT 0200. Ph. (06) 279 8211 Fax (06) 249 2789.Abstracts of all CAEPR Publications can be found at the following WWW address:http://coombs.anu.edu.au/WWWVLPages/AborigPages/CAEPR/caepr-home.html.

CENTRE FOR ABORIGINAL ECONOMIC POLICY RESEARCH(CAEPR)

RECENT DISCUSSION PAPERS

75/1994 'Working for CDEP': a case study of the Community DevelopmentEmployment Projects scheme in Port Lincoln, South Australia, D.E. Smith.

76/1994 Socioeconomic status at the A TSIC regional level, 1986 and 1991: data forregional planning, J.C. Altman and Liu Jin.

77/1994 The relative mobility status of indigenous Australians: setting the researchagenda, J. Taylor and M. Bell.

78/1994 The mobility status of indigenous Australians, J. Taylor and M. Bell.

79/1995 Assessing the relative allocative efficiency of the Native Title Act 1993 andthe Aboriginal Land Rights (Northern Territory) Act 1976, S.L. McKenna.

80/1995 Looking beyond the borderline: development performance and prospects ofSaibai Island, Torres Strait, R. Davis.

81/1995 Performance indicators for Aboriginal Health Services, I. Anderson andM. Brady.

82/1995 Change in the relative economic status of indigenous males in the 1980s:Australia and the United States compared, R.G. Gregory and A.E Daly.

83/1995 Indigenous employment and job segregation in the Northern Territory labourmarket, J. Taylor.

84/1995 Local governments and indigenous Australians: developments and dilemmasin contrasting circumstances, W. Sanders.

85/1995 Mineral development agreements negotiated by Aboriginal communities inthe 1990s, C. O'Faircheallaigh.

86/1995 Negotiations between mining companies and Aboriginal communities:process and structure, C. O'Faircheallaigh.

87/1995 Aboriginal employment, native title and regionalism, J. Finlayson.

88/1995 Native Title Act 1993: implementation issues for resource developers,J.C. Altman.

89/1995 Beyond native title: multiple land use agreements and Aboriginal governancein the Kimberley, P. Sullivan.

90/1995 Australian fiscal federalism and Aboriginal self-government: some issues oftactics and targets, W. Sanders.

91/1995 Enumerating the Aboriginal population of remote Australia: methodologicaland conceptual issues, D.F. Martin and J. Taylor.

92/1995 Twenty years of policy recommendations for indigenous education: overviewand research implications, R.G. Schwab.

93/1995 The economic status of indigenous Australian families, A.E. Daly andD.E. Smith.

94/1995 Equity for Aboriginal families in the 1990s: the challenges for social policy,J. Finlayson.

95/1995 Native title and indigenous Australian utilisation of wildlife: policyperspectives, J.C. Altman, H.J. Bek and L.M. Roach.

96/1995 Change in the relative distribution of indigenous employment by industry,1986-91, L Taylor and Liu Jin.

97/1995 Estimating the private rate of return to education for indigenous Australians,A.E. Daly and Liu Jin.

98/1995 Coping with locational advantage: the economic development potential oftourism at Seisia community, Cape York Peninsula, J.C. Altman.

99/1995 Redfem works: the policy and community challenges of an urban CDEPscheme, D.E. Smith.

100/1995 The calculus of reciprocity: principles and implications of Aboriginalsharing, R.G. Schwab.

101/1995 Money, business and culture: issues for Aboriginal economic policy,D.F. Martin.

102/1995 Indigenous peoples and reshaping Australian institutions: two perspectives,N. Pearson and W. Sanders.

103/1996 Policy implications of rising Aboriginal fertility in the early 1990s,H. Tesfaghiorghis.

104/1996 Change in the relative occupational status of Indigenous workers, 1986-91,J. Taylor and J. Lui.

105/1996 Reforming financial aspects of the Native Title Act 1993: an economicsperspective, J.C. Altman.

106/1996 Indigenous Australians and the socioeconomic status of urbanneighbourhoods, B. Hunter.

107/1996 The comparative economic status of CDEP and non-CDEP communityresidents in the Northern Territory in 1991, J.C. Altman and B. Hunter.

108/1996 Indigenous participation in labour market and training programs,J. Taylor and B. Hunter.

For information on earlier CAEPR Discussion Papers please contact PublicationSales, Centre for Aboriginal Economic Policy Research, Faculty of Arts, AustralianNational University, Canberra ACT 0200. Ph (06) 279 8211 Fax (06) 249 2789.Abstracts of all CAEPR Publications can be found at the following WWW address:http://coombs.anu.edu.au/WWWVLPages/AborigPages/CAEPR/caepr-home.html.